; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt -passes='loop-vectorize,loop-unroll' -force-vector-width=4 -S < %s | FileCheck %s

; Test if the follow-up metadata for loops works fine. The original code is
; something like below. In this case, unrolling should be applied after
; vectorization.
;
; void f(float *a, float x) {
;   #pragma clang loop vectorize(enable) unroll_count(8)
;   for (int i = 0; i < 1024; i++) {
;     a[i] *= x;
;   }
; }
;
define void @f(ptr noundef captures(none) %a, float noundef %x) {
; CHECK-LABEL: define void @f(
; CHECK-SAME: ptr noundef captures(none) [[A:%.*]], float noundef [[X:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x float> poison, float [[X]], i64 0
; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x float> [[BROADCAST_SPLATINSERT]], <4 x float> poison, <4 x i32> zeroinitializer
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX_NEXT_6:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_6]]
; CHECK-NEXT:    [[WIDE_LOAD_7:%.*]] = load <4 x float>, ptr [[TMP14]], align 4
; CHECK-NEXT:    [[TMP15:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_7]]
; CHECK-NEXT:    store <4 x float> [[TMP15]], ptr [[TMP14]], align 4
; CHECK-NEXT:    [[INDEX_NEXT1:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 4
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT1]]
; CHECK-NEXT:    [[WIDE_LOAD_1:%.*]] = load <4 x float>, ptr [[TMP2]], align 4
; CHECK-NEXT:    [[TMP3:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_1]]
; CHECK-NEXT:    store <4 x float> [[TMP3]], ptr [[TMP2]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_1:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 8
; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_1]]
; CHECK-NEXT:    [[WIDE_LOAD_2:%.*]] = load <4 x float>, ptr [[TMP16]], align 4
; CHECK-NEXT:    [[TMP5:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_2]]
; CHECK-NEXT:    store <4 x float> [[TMP5]], ptr [[TMP16]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_2:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 12
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_2]]
; CHECK-NEXT:    [[WIDE_LOAD_3:%.*]] = load <4 x float>, ptr [[TMP6]], align 4
; CHECK-NEXT:    [[TMP7:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_3]]
; CHECK-NEXT:    store <4 x float> [[TMP7]], ptr [[TMP6]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_3:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 16
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_3]]
; CHECK-NEXT:    [[WIDE_LOAD_4:%.*]] = load <4 x float>, ptr [[TMP8]], align 4
; CHECK-NEXT:    [[TMP9:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_4]]
; CHECK-NEXT:    store <4 x float> [[TMP9]], ptr [[TMP8]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_4:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 20
; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_4]]
; CHECK-NEXT:    [[WIDE_LOAD_5:%.*]] = load <4 x float>, ptr [[TMP10]], align 4
; CHECK-NEXT:    [[TMP11:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_5]]
; CHECK-NEXT:    store <4 x float> [[TMP11]], ptr [[TMP10]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_5:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 24
; CHECK-NEXT:    [[TMP12:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_5]]
; CHECK-NEXT:    [[WIDE_LOAD_6:%.*]] = load <4 x float>, ptr [[TMP12]], align 4
; CHECK-NEXT:    [[TMP13:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_6]]
; CHECK-NEXT:    store <4 x float> [[TMP13]], ptr [[TMP12]], align 4
; CHECK-NEXT:    [[INDEX_NEXT_7:%.*]] = add nuw nsw i64 [[INDEX_NEXT_6]], 28
; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[INDEX_NEXT_7]]
; CHECK-NEXT:    [[WIDE_LOAD_8:%.*]] = load <4 x float>, ptr [[TMP17]], align 4
; CHECK-NEXT:    [[TMP18:%.*]] = fmul <4 x float> [[BROADCAST_SPLAT]], [[WIDE_LOAD_8]]
; CHECK-NEXT:    store <4 x float> [[TMP18]], ptr [[TMP17]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw nsw i64 [[INDEX_NEXT_6]], 32
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1024
; CHECK-NEXT:    br i1 [[TMP4]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[SCALAR_PH]]:
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 0, %[[ENTRY]] ]
; CHECK-NEXT:    br label %[[FOR_BODY:.*]]
; CHECK:       [[FOR_BODY]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT_7:%.*]], %[[FOR_BODY]] ]
; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV]]
; CHECK-NEXT:    [[LOAD:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; CHECK-NEXT:    [[MUL:%.*]] = fmul float [[X]], [[LOAD]]
; CHECK-NEXT:    store float [[MUL]], ptr [[ARRAYIDX]], align 4
; CHECK-NEXT:    [[IV_NEXT:%.*]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[ARRAYIDX_1:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT]]
; CHECK-NEXT:    [[LOAD_1:%.*]] = load float, ptr [[ARRAYIDX_1]], align 4
; CHECK-NEXT:    [[MUL_1:%.*]] = fmul float [[X]], [[LOAD_1]]
; CHECK-NEXT:    store float [[MUL_1]], ptr [[ARRAYIDX_1]], align 4
; CHECK-NEXT:    [[IV_NEXT_1:%.*]] = add nuw nsw i64 [[IV]], 2
; CHECK-NEXT:    [[ARRAYIDX_2:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_1]]
; CHECK-NEXT:    [[LOAD_2:%.*]] = load float, ptr [[ARRAYIDX_2]], align 4
; CHECK-NEXT:    [[MUL_2:%.*]] = fmul float [[X]], [[LOAD_2]]
; CHECK-NEXT:    store float [[MUL_2]], ptr [[ARRAYIDX_2]], align 4
; CHECK-NEXT:    [[IV_NEXT_2:%.*]] = add nuw nsw i64 [[IV]], 3
; CHECK-NEXT:    [[ARRAYIDX_3:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_2]]
; CHECK-NEXT:    [[LOAD_3:%.*]] = load float, ptr [[ARRAYIDX_3]], align 4
; CHECK-NEXT:    [[MUL_3:%.*]] = fmul float [[X]], [[LOAD_3]]
; CHECK-NEXT:    store float [[MUL_3]], ptr [[ARRAYIDX_3]], align 4
; CHECK-NEXT:    [[IV_NEXT_3:%.*]] = add nuw nsw i64 [[IV]], 4
; CHECK-NEXT:    [[ARRAYIDX_4:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_3]]
; CHECK-NEXT:    [[LOAD_4:%.*]] = load float, ptr [[ARRAYIDX_4]], align 4
; CHECK-NEXT:    [[MUL_4:%.*]] = fmul float [[X]], [[LOAD_4]]
; CHECK-NEXT:    store float [[MUL_4]], ptr [[ARRAYIDX_4]], align 4
; CHECK-NEXT:    [[IV_NEXT_4:%.*]] = add nuw nsw i64 [[IV]], 5
; CHECK-NEXT:    [[ARRAYIDX_5:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_4]]
; CHECK-NEXT:    [[LOAD_5:%.*]] = load float, ptr [[ARRAYIDX_5]], align 4
; CHECK-NEXT:    [[MUL_5:%.*]] = fmul float [[X]], [[LOAD_5]]
; CHECK-NEXT:    store float [[MUL_5]], ptr [[ARRAYIDX_5]], align 4
; CHECK-NEXT:    [[IV_NEXT_5:%.*]] = add nuw nsw i64 [[IV]], 6
; CHECK-NEXT:    [[ARRAYIDX_6:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_5]]
; CHECK-NEXT:    [[LOAD_6:%.*]] = load float, ptr [[ARRAYIDX_6]], align 4
; CHECK-NEXT:    [[MUL_6:%.*]] = fmul float [[X]], [[LOAD_6]]
; CHECK-NEXT:    store float [[MUL_6]], ptr [[ARRAYIDX_6]], align 4
; CHECK-NEXT:    [[IV_NEXT_6:%.*]] = add nuw nsw i64 [[IV]], 7
; CHECK-NEXT:    [[ARRAYIDX_7:%.*]] = getelementptr inbounds nuw float, ptr [[A]], i64 [[IV_NEXT_6]]
; CHECK-NEXT:    [[LOAD_7:%.*]] = load float, ptr [[ARRAYIDX_7]], align 4
; CHECK-NEXT:    [[MUL_7:%.*]] = fmul float [[X]], [[LOAD_7]]
; CHECK-NEXT:    store float [[MUL_7]], ptr [[ARRAYIDX_7]], align 4
; CHECK-NEXT:    [[IV_NEXT_7]] = add nuw nsw i64 [[IV]], 8
; CHECK-NEXT:    br i1 true, label %[[EXIT_LOOPEXIT:.*]], label %[[FOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       [[EXIT_LOOPEXIT]]:
; CHECK-NEXT:    br label %[[EXIT]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %for.body

for.body:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
  %arrayidx = getelementptr inbounds nuw float, ptr %a, i64 %iv
  %load = load float, ptr %arrayidx, align 4
  %mul = fmul float %x, %load
  store float %mul, ptr %arrayidx, align 4
  %iv.next = add nuw nsw i64 %iv, 1
  %comp = icmp eq i64 %iv.next, 1024
  br i1 %comp, label %exit, label %for.body, !llvm.loop !0

exit:
  ret void
}

!0 = distinct !{!0, !1, !2}
!1 = !{!"llvm.loop.vectorize.enable", i1 true}
!2 = !{!"llvm.loop.vectorize.followup_all", !3, !4}
!3 = !{!"llvm.loop.isvectorized"}
!4 = !{!"llvm.loop.unroll.count", i32 8}
;.
; CHECK: [[LOOP0]] = distinct !{[[LOOP0]], [[META1:![0-9]+]], [[META2:![0-9]+]]}
; CHECK: [[META1]] = !{!"llvm.loop.isvectorized"}
; CHECK: [[META2]] = !{!"llvm.loop.unroll.disable"}
; CHECK: [[LOOP3]] = distinct !{[[LOOP3]], [[META1]], [[META2]]}
;.